Science Arts & Métiers (SAM)
is an open access repository that collects the work of Arts et Métiers Institute of
Technology researchers and makes it freely available over the web where possible.
This is an author-deposited version published in: https://sam.ensam.eu
Handle ID: .http://hdl.handle.net/10985/11963
To cite this version :
Djahida SIDANE, Hafit KHIREDDINE, Fatima BIR, Sabeha YALA, Alex MONTAGNE, Didier CHICOT HydroxyapatiteTiO2SiO2Coated 316L Stainless Steel for Biomedical Application -Metallurgical and Materials Transactions A - Vol. 48, n°7, p.3570–3582 - 2017
Any correspondence concerning this service should be sent to the repository Administrator : archiveouverte@ensam.eu
Science Arts & Métiers (SAM)
is an open access repository that collects the work of Arts et Métiers ParisTech
researchers and makes it freely available over the web where possible.
This is an author-deposited version published in: http://sam.ensam.eu
Handle ID: .http://hdl.handle.net/null
To cite this version :
DJAHIDA SIDANE, HAFIT KHIREDDINE, FATIMA BIR, SABEHA YALA, Alex MONTAGNE, Didier CHICOT - Hydroxyapatite-TiO2-SiO2-Coated 316L Stainless Steel for Biomedical Application - Metallurgical and Materials Transactions A - Vol. 48, n°7, p.3570–3582 - 2017
Hydroxyapatite-TiO
2
-SiO
2
-Coated 316L Stainless
Steel for Biomedical Application
DJAHIDA SIDANE, HAFIT KHIREDDINE, FATIMA BIR, SABEHA YALA, ALEX MONTAGNE, and DIDIER CHICOT
This study investigated the effectiveness of titania (TiO2) as a reinforcing phase in the hydroxyapatite (HAP) coating and silica (SiO2) single layer as a bond coat between the TiO2-reinforced hydroxyapatite (TiO2/HAP) top layer and 316L stainless steel (316L SS) substrate on the corrosion resistance and mechanical properties of the underlying 316L SS metallic implant. Single layer of SiO2film was first deposited on 316L SS substrate and studied separately. Water contact angle measurements, X-ray photoelectron spectroscopy, and Fourier transform infrared spectrophotometer analysis were used to evaluate the hydroxyl group reactivity at the SiO2outer surface. The microstructural and morphological results showed that the reinforcement of HAP coating with TiO2 and SiO2 reduced the crystallite size and the roughness surface. Indeed, the deposition of 50 vol pct TiO2-reinforced hydroxyapatite layer enhanced the hardness and the elastic modulus of the HAP coating, and the introduction of SiO2inner layer on the surface of the 316L SS allowed the improvement of the bonding strength and the corrosion resistance as confirmed by scratch studies, nanoindentation, and cyclic voltammetry tests.
DOI: 10.1007/s11661-017-4108-8
The Minerals, Metals & Materials Society and ASM International 2017
I. INTRODUCTION
H
YDROXYAPATITE (Ca10(PO4)6(OH)2, HAP) isfound to be the preferred bioactive ceramic due to its chemical, structural, and biological similarity to the inorganic component of human bones and to its direct bonding capability to surrounding tissues.[1] However, the mechanical weakness of the HAP limits its practical applications to those requiring little or no load-bearing locations. Therefore, to widen the applicability of the HAP to the sites bearing substantial load, such as dental or hip implant, a coating system including HAP on the metallic implant is being used. Metallic materials such as 316L stainless steel (316L SS), Co-Cr alloy, Ti, and Ti-6Al-4V alloy are widely used biomaterials for hard tissue replacement due to their superior tensile strength, fracture toughness, corrosion resistance, and biocom-patibility with the internal human environment.[2–5] 316L SS is used in orthopaedic medical fields due to its unique property of biocompatibility, cost
effectiveness, and corrosion resistance.[6,7] It is shown that the introduction of an intermediate thin layer of titania (TiO2), which possesses combined advantages of
biocompatibility and corrosion resistance properties,[8,9] between the metal substrate and the HAP layer signif-icantly improves the mechanical properties and corro-sion resistance of the HAP coatings.[10–14]On the other side, a large number of articles have reported that the addition of TiO2as secondary phase into HAP coatings
improved the mechanical and bonding strength of the HAP coatings.[15–17] Moreover, it is recognized that silica (SiO2)-based sol–gel systems usually have a high
content of surface silanol groups, which have been reported to promote in vitro and in vivo nucleation of apatite.[18,19] Galliano et al.[20] reported that the silica films deposited on 316L SS substrate were able to reduce both the corrosion attack on the steel and the iron diffusion to the sample surface. However, not much study has been reported on the effect of silica-coated 316L SS on the structure, mechanical, and electrochem-ical behavior of the TiO2-reinforced hydroxyapatite
coatings (TiO2/HAP). Therefore, the objective of the
current study is to investigate the hydrophilic properties of SiO2film, which can improve the bonding capability
of the TiO2/HAP layer with the substrate and provide a
first barrier against corrosion of the 316L SS in simulated human body fluids. Further, nanoindentation test is performed using the continuous stiffness mea-surement (CSM) mode in order to determine the
variation of the hardness and elastic modulus of the coatings as a function of the indenter displacement.
Finally, the addition of titania phase into HAP and the use of silica as a bonding oxide layer between the TiO2/HAP top layer and the substrate are studied within
the objective to improve the mechanical properties, bonding strength, and corrosion resistance of the hydroxyapatite-TiO2-SiO2-coated 316L SS system.
II. MATERIALS AND METHODS
A. Experimental Details
Tetraethyl orthosilicate (TEOS, Fluka 98 pct) was mixed with methyltriethoxysilane (MTES, Fluka 98 pct) in equimolar ratio. Hydrochloric acid (0.1 N) and acetic acid (HOAc) were added as catalysts, and their role is to increase the hydrolysis and condensation. The reaction rate to the concentrations of the chemical species present in the reaction mixture (H2O+HOAc)/
(TEOS+MTES) was 1.7.[20] After stirring the solution under room temperature conditions for 24 hours, a transparent viscous sol was obtained. Precursor solution for HAP coating was prepared by dissolving phospho-rous pentoxide (P2O5, Prolabo 100 pct) (0.5 mol/l) and
calcium nitrate tetrahydrate (Ca(NO3)2Æ4H2O, Fluka 98
pct) (1.67 mol/l) in absolute ethanol,[21]and then the two solutions were mixed and continuously stirred for 24 hours at room temperature. Titanium isopropoxide (TIP, Fluka 100 pct) was used as a titania precursor. The reactivity towards water is modified by acetic acid (molar ratio of TIP/HOAc = 1/10) which is also used as catalyst. 2-methoxy ethanol was added to adjust the viscosity of the solution.[22]This solution was vigorously stirred under room temperature conditions. For the preparation of the composite sols, HAP and TiO2
solutions were mixed with different volumes: 0, 20, and 50 vol pct TiO2, and then the solutions were
continu-ously stirred for 14 hours.
The 316L SS used as the substrates with dimensions of 20 9 10 9 5 mm was mechanically polished using different silicon carbide grit papers from 120 to 1200 grades. A mirror polishing was done using diamond paste of 2 and of 0.7 lm in the final step. The substrates were ultrasonically degreased with acetone and washed with running double-distilled water, and then they were dried at 423 K (150 C) for 10 minutes.
Finally, the substrates were immersed in the HAP and TiO2/HAP sols and dipped in the suspensions at a speed
of 80 mm/min, and then dried and annealed at 773 K (500 C) for 60 minutes. The obtained coatings are referenced as follows: 20 vol pct TiO2(H80T20), 50 vol
pct TiO2 (H50T50). The SiO2layer was coated on the
surface of the substrate at a speed of 100 mm/min, and then dried and annealed at 773 K (500 C) for 60 minutes. The pure HAP and composite HAP/TiO2sols
were subsequently dip-coated on the outer surface of the SiO2film at a speed of 80 mm/min and annealed at 773
K (500 C) for 60 minutes. The obtained coatings are referenced as follows: 0 vol pct TiO2(HAP-SiO2), 20 vol
pct TiO2(H80T20-S), and 50 vol pct TiO2(H50T50-S).
In addition, pure TiO2film was prepared by dipping at a
speed of 20 mm/min, and then dried and annealed at 723 K (450C) for 60 minutes.[14]
B. Characterizations
The chemical states of the coatings were characterized by X-ray photoelectron spectroscopy (XPS) with a Thermo L-alpha spectrometer, using monochromatized Al-Ka radiation as the excitation source (hm =1486.6 eV), collected at 0 deg from the surface normal, and detected with a hemispherical analyzer. The spot size of the XPS source on the sample was about 200 lm, and the pass energy was set of 20 eV. During data acqui-sition pressure was kept below 1 9 109 Torr. Spectra were fitted using a 10 pct linear combination of Gaussian and Lorentzian profiles. Peak positions obtained after analysis were found essentially constant (±0.3 eV).
Water contact angles were obtained with a Digidrop Contact Angle Meter (GBX Surface Science Technolo-gies). Measurements were carried out with 3 ll drops of ultra-pure water under ambient atmospheric conditions. Three drops were applied on each sample at different locations, to assure reproducibility and reliability of results.
Both the coating thickness and roughness were measured using a profilometry analysis ‘‘DEKTAK 150 SURFACE PROFILERT’’ by recording the surface profile of coated and uncoated regions over a single measurement run. The surface of the coating was scanned at an interval of 1000 to 8000 lm. Three different areas were scanned and measured to determine a mean value for the thickness and the roughness parameters.
The surface morphology of the samples was observed
with a scanning electron microscope (SEM)
(JSM-5400LV, JEOL) operating at 0 to 30 kV, associ-ated with an ultra-thin window Si(Li) detector for the energy-dispersive X-ray measurements (EDX) (GEN-ESIS, Eloı¨se SARL).
The different phases of the coatings were analyzed by X-ray diffraction Expert Prof Panalytical type MPD/ system vertical h/h, using radiation source (Cu-Ka = 1.5406 A˚) operating at 40 kV and 30 mA. The XRD diffraction patterns were collected over a 2h range located between 20 and 80 deg using an incremental step size of 0.02 deg with 6 seconds of acquisition time per Table I. The Composition of SBF Solution at 310 K (37 °C)
Order Reagent Amount
1 NaCl 7.996 g 2 NaHCO3 0.350 g 3 KCl 0.224 g 4 K2HPO4.3H2O 0.228 g 5 MgCl2Æ6H2O 0.305 g 6 CaCl2 0.278 g 7 Na2SO4 0.071 g 8 (CH2OH)3CNH2(Tris) 6.057 g
9 HCl (1M) appropriate amount for
step. The phase identification was performed by com-paring the experimental XRD patterns to standards compiled by the International Center for Diffraction Data-Powder Diffraction Files (ICDD-PDF). Data were
treated with Software X’Pert HighScore. The HAP and TiO2/HAP average crystallite sizes (L) were estimated
with broadening XRD peaks using Debye–Scherrer equation:
Fig. 1—(a) Contact angle between water, SiO2oxide film, and 316L stainless steel substrate, (b) Images of water droplets on the surfaces of the substrate and SiO2film obtained after 20 s (c) FTIR spectrum of the single SiO2film, (d) XPS spectrum of the single SiO2film, (e, f) The spectra of O1s and Si2p, respectively.
Fig. 2—SEM observations corresponding to (a) H80T20, (b) H50T50, (c) H80T20-S, and (d) H50T50-S coatings, (e) 316L SS sub-strate-SiO2-H50T50 interfaces, (f) EDX elemental analysis.
Table II. Structural Parameters of the Uncoated and Coated 316L SS
Sample Thickness (nm) Roughness (nm)
316L SS — 20 HAP 1600 1310 HAP-SiO2 1610 921 H80T20 1336 740 H80T20-S 1544 642 H50T50 1110 681 H50T50-S 1310 550 Lhkl¼ Kk b cos h; ½1
where k is the wavelength of the Cu-Ka radiation, b is full width at middle height (the Bragg peak maximum intensity (deg)), h is Bragg’s angle (deg), and K is the Scherrer constant (equal to 0.9 when the width is measured at middle height of the diffraction peak). The mean crystallite size for pure HAP sample was calcu-lated from the (2 1 1) reflection peak.
The structural analysis was carried out using the Fourier transform infrared (FTIR) spectrophotometer
instrument (IRAffinity-1, SHIMADZU). FTIR spectra were recorded in the range of 400 to 4000 cm1with a resolution of 4 cm1. The deposited films were scraped off as powders from the substrate and mixed with KBr powder (80 pct in weight), to form an infrared trans-parent pellet.
Corrosion behavior of the samples was evaluated by potentiodynamic cyclic voltammetry tests by Voltalab (Serial: 913V708/INT), interfaced with a computer, and loaded with VoltaMaster 4 software in simulated human body fluid (SBF) at 310 ± 274 K (37 ± 1C). The SBF solution was prepared using Kokubo and Takadama’s formulation[23] by dissolving reagent-grade NaCl, NaHCO3, KCl, K2HPO4Æ3H2O, MgCl2Æ6H2O, CaCl2,
and Na2SO4into distilled water (1 L) and buffered at pH
7.40 with (CH2OH)3CNH2 and 1M HCl solution
(Table I). Before conducting the corrosion studies, the specimens were immersed in SBF solution for 1 hour in order to stabilize the system.[14] Moreover, a renewed solution was used for each experiment. The exposed area of the samples in the SBF solution was 1 cm2. A platinum electrode was used as the auxiliary electrode, and the saturated calomel electrode (SCE) was used as the reference electrode. Corrosion potential (ECorr) and
corrosion current density (iCorr) were determined using
the Tafel diagram with sweeping potential from1000 to +1000 mV at the rate of 1 mVs1, and at least three similar results were required to ensure reproducibility.
The adhesion of the coating was estimated using a scratch tester (Millennium in accord with Standard ISO/ EN 1071-3) with a spherical Rockwell C diamond indenter of 200 lm in radius. The scratch tests were performed on the coating by applying the load, which increased monotonously at the loading rate of 10000 mN/min while the specimen was shift at the constant speed of 1500 lm/min. The applied force immediately started to increase linearly with time. These conditions lead to a total scratch length of 1500 lm. The load at which coating was removed from the substrate is referred as the critical load (Lc). The scratch track was
observed using optical microscope. Five tests for each sample were recorded.
Nanoindentation experiments were performed with a Nano Indenter XP (MTS Nano Instruments) employ-ing a Berkovich diamond indenter calibrated usemploy-ing a reference sample of known modulus (fused quartz, E= 72 GPa). The samples were fixed on a metallic support using the heat softening glue crystalbond 509. A regular array of 5 9 5 indentation tests has been performed at the surface of the different coated mate-rials in order to obtain representative variations of the hardness and the elastic modulus as a function of the indenter displacement. The maximum indentation depth reached by the indenter was fixed at 2000 nm and the strain rate was set constant and equal to 0.05 s1. The instrument was operated in the continuous stiffness measurement (CSM) mode allowing the computation of the elastic modulus and the hardness continuously during the indentation loading. The harmonic displace-ment was 2000 nm and the frequency was 45 Hz. The elastic modulus of the coating, EC, is deduced from the
reduced modulus, ERC, given by the instrument, which
takes into account the elastic properties, Ei and mi,
related to the indenter material and the Poisson’s ratio of the coating, mc: EC¼ 1 v2C 1 ERC þ 1 v 2 i Ei 1 : ½2
For a diamond indenter, the elastic modulus, Ei, and the
Poisson’s coefficient, mi, are equal to 1140 GPa and 0.07,
respectively.[24]mcis taken equal to the mean value of 0.3
because the analysis deals with a multilayered coating.
III. RESULTS AND DISCUSSION
A. Hydrophilic Properties of SiO2Surface
The effect of the SiO2 film on the hydrophilic
properties of the 316L SS samples was evaluated by measuring their water contact angles (Figures1(a) and (b)). Contact angle variations are compared to those measured on the as-prepared 316L SS substrate, i.e., polished and degreased, then dried at 423 K (150 C), and those measured on the SiO2 single-layer-coated
316L SS. All the specimens were aged for 60 days in the ambient environment. As shown in Figure1(a), the contact angle does not vary with time indicating that the shape and/or the size of the water drop applied on the supports did not vary with time during the measure-ment. The representative images of the water droplets deposited on the substrate and SiO2 film surfaces are
Fig. 3—XRD patterns of the (a) HAP, (b) H80T20, (c) H50T50, and (d) H50T50-S coatings. (*) Hydroxyapatite, (S) Substrate, and (e) Anatase.
Table III. Crystalline and Lattice Parameters of HAP and TiO2/HAP Coatings
HAP H80T20 H50T50
a (A˚) 9.410 9.394 9.392
c (A˚) 6.887 6.884 6.870
L211(nm) 135 61.7 30.8
presented in Figure1b. The contact angle measurements indicate that the substrate is partially wetted by water and that wettability of the steel is the lowest, the 316L steel substrate has contact angle of 79.8 ± 0.1 deg, water wets silica rather well than the substrate, and it represents a contact angle of 34.2 ± 0.02 deg. These results are in agreement with the previous works of Houmard et al.[25]and Permpoon et al.[26]showing that silica films prepared by sol–gel method and heat-treated at 773 K (500 C) continuously exhibited a very slow contact angle increase over a period of 60 days aging, which confirms the natural hydrophilicity of a silica surface due to the hydroxyl content on the film surface. FTIR analyses performed on the same films after 60 days aging are represented in Figure1(c). There is a broad band in the range of 3200 to 3800 cm1 corresponding to stretching vibration of different hydroxyl groups associated to absorbed free water and to Si-OH (silanols) groups linked to molecular water through hydrogen bonds as well as isolated free surface silanols. The bands around 1556 and 1656 cm1 correspond to bending vibrations of OH bonds of water molecules. The band around 565 cm1 is attributed to Si-O bonds in an amorphous phase. The bands at 460, 800 cm1, and in the range of 1000-1220 cm1 are attributed to the bond bending and bond stretching vibrations of the Si-O-Si units in silica. The band at 935 cm1 is assigned to the stretching vibration of a small amount of Si-OH groups.[27–29]
Figure1(d) related to the XPS spectrum shows the elemental composition of deposited SiO2 film. The
binding energy peaks are corresponding to Si2s, Si2p, and O1s of SiO2film. It is inferred that the C1s peak at
284.7 eV essentially reflects the amount of carbon contamination at the outer surface. Previous stud-ies[26,30] have shown that alkoxy groups, which might contribute to the C1s peak, are not present in SiO2film
heated at 500 C, and alkoxy groups arising from the silica precursors are completely decomposed after annealing at 773 K (500C). In addition, XPS analysis
was used to investigate the surface hydrophilic proper-ties of SiO2film. The presence of OH groups on the SiO2
outer surface is studied from the deconvolution of the O1s peak. The O1s peak could be decomposed in two components, i.e., Si-OH and Si-O components, using Lorentzian/Gaussian functions (Figures1(e) and (f)). The O-H component essentially traduces the presence of surface hydroxyl groups; the O1s region includes two peaks (Figure1(e)). One component of the O1s peak is attributed to the Si-O (533 eV) and the other one is assigned to the hydroxyl group Si-OH (535 eV). Binding energies of these components were subsequently com-pared to those of component measured for pure SiO2
film.[26] The hydroxyl content (pct) is the ratio of the area of 535 eV component to the total area of the two O1s components. The hydroxyl content for SiO2 film
was calculated as 4 pct. This observation is in accor-dance with FTIR results, which depicted an amount of surface OH groups for the SiO2layer. The
deconvolu-tion of the Si2p spectrum into a single Si4+component located at 103.5 eV is presented in Figure1(f), and the location of this component is very similar to that measured for pure SiO2film.[31]
B. Morphological and Structural Characterization of Sol–Gel-Derived Hydroxyapatite Coatings
Figure2 shows the surface of the HAP/TiO2
depos-ited on the SiO2-coated and uncoated 316L SS
sub-strates. The images show highly intermixed composite phases. From Figure2(c) and (d), it can be seen that H80T20-S and H50T50-S have the continuous and regular microstructure. Moreover, the deposits cover entire surface of the substrate compared to H80T20 (Figure2(a)) and H50T50 (Figure2(b)). Figure2(e) illustrates the SEM micrograph of the H50T50 top layer/SiO2inner layer/substrate interfaces. Some
micro-cracks are generated in the coating, which can represent the interface separating the H50T50 and SiO2 layers;
otherwise the deposit is mostly uniform throughout the rest areas, and no delamination was observed. The EDX spectrum obtained on the H50T50-S coating indicates the peaks corresponding to the elements that are originating from the substrate, HAP, TiO2, and SiO2
phases; in addition, carbon element was detected. The thickness and roughness of the sol–gel-derived coatings are collected in TableII. All the coatings presented a thickness less than 2 lm. Pure HAP coating exhibited a higher surface roughness (.1310 nm). The addition of both TiO2and SiO2decreased the roughness
of the coatings up to 550 nm. According to Sidane et al.,[32] the SEM surface examination of the HAP coating exhibited a porous surface composed of spher-ical agglomerates. After the addition of TiO2into HAP
phase, it can be seen that H80T20 (Figure2(a)) and H50T50 (Figure2(b)) composite coatings have regular surface. The porosity and the agglomerated particles are reduced and almost disappeared after the insertion of SiO2inner layer between the composite top layer and the
substrate, so the resulted H80T20-S and H50T50-S bilayer coatings have the uniform surface as it is shown in Figures2(c) and (d).
Fig. 4—FTIR spectra of the (a) H50T50, (b) H50T50-S coatings, and (c) TiO2film.
The composition and microstructure of the coatings were characterized by XRD analysis (Figure3). Accord-ing to the International Center for Diffraction Data (ICDD) patterns, the pattern of the HAP coating (Figure3(a)) shows peak positions, which correspond to a crystallized HAP structure (PDF no. 09-0432). The HAP characteristic triplet peaks at (211), (112), and (300) planes are observed for 2h values between 31 and 33 deg, and the peak at the plane (002) is observed for 2h = 25.88 deg. In addition to the existing HAP peaks, diffraction peaks corresponding to steel substrate (PDF no. 00-006-0694) are developed. Addition of 20 vol pct TiO2 (Figure3(b)) and 50 vol pct TiO2 (Figure3(c))
shows almost similar diffraction pattern such as pure HAP, but the peaks become wider with a change in the position and a decrease in the intensity. The mean peak corresponding to the crystalline structure of anatase (2h = 25.35 deg) (PDF no. 00-004-0477) is confused with that of HAP (2h = 25.88 deg). When TiO2
concentration increases further, a peak shift and the decrease in the XRD peak intensity are observed. This confirms the previous results of Nathanael et al.[15,16] who attributed the initial small decrease in the intensities of the HAP peaks to the small inclusion of TiO2 in
HAP. There are different proposals in the literature for possible Ti substitution mechanisms in HAP structure.
According to Ergun et al.[33] and Riberio et al.,[34] the ionic radius of Ti4+(0.68 A˚) is much smaller than the ionic radius of Ca2+(0.99 A˚), and then the substitution of Ca by low concentrations of Ti can be occurred and results in a decrease in the cell lattice parameters and crystal domain size. Therefore, lattice disorder increased with the increasing of Ti content. Thus, this lattice disorder greatly inhibits the crystallization and makes it difficult to obtain high crystallinity in HAP-modified structure.[35] Crystallite size and lattice parameters, calculated from diffraction line broadening, are regrouped in TableIII. They decrease as the content of TiO2 increases from 20 to 50 vol pct, respectively.
After the deposition of SiO2inner layer on the 316L SS
substrate, the pattern of H50T50-S (Figure3(d)) reveals an important shift and reduction in the intensity of HAP peaks. This behavior could be related to a combination between the SiO2, TiO2, and HAP phases through
chemical bonding. There is no crystalline phase of SiO2
detected by the X-rays diffraction because the microstructure of the SiO2is amorphous.[32]
Figure4 shows the FTIR spectra of the H50T50, H50T50-S, and pure TiO2coatings. FTIR spectrum of
H50T50 (Figure4(a)) exhibits the well-defined bands associated with the presence of PO4groups at 473, 563,
600, 964, 1049, and 1089 cm1. The broad and
Fig. 5—Potentiodynamic curves of (a) the uncoated 316L stainless steel substrate (b) H50T50, (c) H50T50-S, (d) HAP-SiO2, (e) H80T20, (f) H80T20-S, and (g) HAP coatings in SBF solution at 310 K (37C).
Table IV. Corrosion and Scratch Test Parameters of the Uncoated and Coated 316L SS
Sample
Corrosion Test Scratch Test
ECorr(mV) iCorr(lA cm2) Scratch length (lm) Critical load, Lc (mN)
316L SS - 690 1.120 — — HAP - 484 0.849 — — HAP-SiO2 - 445 0.812 — — H80T20 - 423 0.801 820 4820 H80T20-S - 383 0.724 880 4980 H50T50 - 395 0.758 900 5050 H50T50-S - 362 0.251 980 5680
high-intensity band extending from 2500 to 3600 cm1 and the band at 1632 cm1 correspond to the hydro-gen-bonded H2O molecules. The bands of CO32-groups
are indicated as a singlet at 873 cm1and in the range of
1400 to 1540 cm1. The origin of the carbonate bands is due to the absorption of atmospheric CO2 into the
ethanol solution.[14,36] In addition, the FTIR spectrum shows the vibrational bands related to the stretching vibration of Ti-O at 467 and 565 cm1 and the asymmetric broad of Ti-O-Ti at about 810 cm1in the anatase phase of TiO2 lattice, as shown in
Figure4(c).[27] An intensity decrease of the OH bands at 3572 and 630 cm1 arising from the stretching and vibrational modes, respectively, of the OH- ions in the hydroxyapatite structure is noticed. These observations are in agreement with the results of XRD patterns. The librational mode of the OH group (632 cm1) is especially sensitive to substitutions in the apatite struc-ture. The strength of this peak is well known to correlate with the degree of crystallinity of HAP. In the presence of Ti, the area of the OH librational band decreases, suggesting a decrease in HAP crystallinity, and conse-quently a decrease in crystallite size. The decrease in intensity of the OH peak at 3572 cm1is also indicative of changes in the HAP structure.[34] After insertion of SiO2 inner layer, the FTIR spectrum of H50T50-S
(Figure4(b)) exhibits the bands corresponding with the presence of hydroxyapatite but their resolution decreases. The spectrum exhibits the broadening of absorption bands corresponding to phosphates, hydrox-yls, and carbonates with the presence of well-developed bands related to Ti-O and Ti-O-Ti bonds. Indeed, the development of a weaker absorption peak of Ti-O-Si
Fig. 6—(A) Optical images representative of the scratch length realized on the surface of (a, b) H80T20, (c, d) H80T20-S, (e, f) H50T50, and (g through i) H50T50-S coatings; load is progressively increasing from right to left; (a), (c), (e), and (g) are related to the initial scratch. (B) Load–displacement graph corresponding to the scratch track presented in the (b, d, f, and h) optical images.
Fig. 7—Schematic diagrams revealing possible bonding mechanism of TiO2/HAP phases on the surface of SiO2 film during the immer-sion in mixed TiO2/HAP sol, through hydroxyl group association.
Fig. 8—Hardness vs the indenter displacement performed on H80T20, H50T50, and H50T50-S coatings.
Table V. The Elastic Modulus and the Fitting Parameters of the TIO2/HAP Coatings
Perriot and Barthel Avrami
EF (GPa) ES (GPa) x0 n EF (GPa) ES (GPa) ks ns H80T20 19 176 3.5 2.24 21 177 21 2 H50T50 32 183 5 2.9 33 188 10.4 2.6
Fig. 9—Models applied on the indentation data obtained from nanoindentation test performed on (a, b) H80T20, (c, d) H50T50, and (e, f) H50T50-S coatings.
mixed bond at 925 cm1 was observed,[27] which demonstrates the Ti-O-Si bond formation due to the bonding of Si-OH groups present on the surface of the SiO2 film underlying layer as shown in the previous
section.
C. Corrosion Resistance
Figure5 shows the potentiodynamic polarization plots of the coated and uncoated 316L SS substrate specimens immersed in SBF solution at 310 ± 274 K (37 ± 1 C). Results of electrochemical tests revealed the influence of the different deposits on the corrosion resistance of the 316L steel. The corrosion parameters determined from these curves by means of Tafel extrapolation method are summarized in Table IV. The corrosion parameters are corrosion potential (ECorr)
and corrosion current density (iCorr). For H50T50-S
(Figure 5(c)) and H80T20-S (Figure5(f)), the potentio-dynamic curves were shifted to the right when compared to H50T50 (Figure5(b)) and H80T20 (Figure5(e)) curves, respectively. Representative potentiodynamic curve obtained for HAP-SiO2(Figure5(d)) was shifted
to the positive potential side compared to pure HAP (Figure 5(g)) and 316L SS sample (Figure5(a)), respec-tively. The corrosion resistance of the specimens increases when decreasing iCorr. For the substrate, iCorr
was about 1.12 lA cm2, and then decreased to 0.849 lA cm2for HAP and to 0.812 lA cm2for HAP-SiO2
bilayer coating. The iCorrof H80T20-S specimens which
is about 0.724 lA cm2 indicates that they are more corrosion resistant than H50T50 (. 0.758 lA cm2) and H80T20 (. 0.801 lA cm2), which are in turn more resistant than HAP-SiO2. When compared to other
specimens, H50T50-S exhibits a lower corrosion current (. 0.251 lA cm2). The noble behavior of the coated 316L SS samples can be attributed to the denser nature of the coating surface, due to the formation of apatite precipitates resulting from the reaction between the calcium and phosphate ions in the SBF solution. Indeed, the presence of the SiO2 inner layer prevents the
electrolyte to infiltrate into the deeper portion of the coating through the pores and cracks existing in the coating causing pitting corrosion. A passive current plateau is noticed on the potentiodynamic polarization curves recorded on HAP-SiO2 (Figure5(d)), H80T20
(Figure 5(e)), and H80T20-S (Figure5(f)). The passiva-tion behavior indicates that protective surface films have been formed on the surface of the specimens exposed to the SBF solution. Potentiodynamic curves recorded on H50T50 (Figure5(b)) and H50T50-S (Figure5(c)) are quite similar. The anodic current density increased with increasing potential suggesting the formation of a thin oxide film on the surface of the coatings. The further increase of the anodic current density suggests a degradation of that passive film.[37,38] Thus, the rein-forcement of the HAP coating with TiO2 and SiO2
improved the corrosion resistance of the 316L steel substrate. It is noted that the increase of TiO2
concen-tration in the hydroxyapatite phase considerably increased the uniformity of the coatings and thereby decreased the surface roughness.[15] After the insertion
of SiO2 inner layer on 316L SS substrate, H50T50-S
bilayer coating showed a smooth surface, (see TableII; Figure2) and exhibited a higher corrosion resistance. The significantly reduced corrosion current density in the H50T50-S (. 0.251 lA cm2) clearly demonstrates the improvement obtained in the corrosion resistance. The surface roughness and surface morphology can considerably change corrosion and corrosion rate. A higher corrosion resistance is obtained for surfaces with lower roughnesses.[39]
D. Bonding Strength
Figure6A shows the optical microscopy images corresponding to the scratch track realized on the surface of coatings during the scratch force test. Load is progressively increasing from right to left. According to Figure6A, no visible wear debris were come out from the coatings in the initial stage of small-applied load (Figure6A(a, c, e, g)). As the load was increased, the cracks appeared on the surface and became more severe until the delamination of the coating from the substrate (Figure6A(b, d, f, i)). The corresponding load is recorded as the critical load Lc. No important damage is inspected on the scratching surface corresponding to H50T50-S coating (Figure 6A(h)), the failure was occurred only once the load is increased as it is shown in Figure 6A(i). The critical loads (Lc1, Lc2, Lc3, and
Lc4) at which the coating was removed from the
substrate are indicated with white arrow in Figure6A(b, d, f, i), respectively. They are determined from the curves illustrated in Figure6B which presents load as a function of the scratch length (displacement) and are regrouped in TableIV. From TableIV, it can be observed that the insertion of the SiO2 inner layer
improved the bonding strength of the H80T20 compos-ite coating to the substrate. H80T20 presents approxi-mately a critical load (Lc1) of 4820 mN, while H80T20-S
presents approximately a critical load (Lc2) of 4980 mN.
Addition of 50 vol pct TiO2to HAP also improves the
bonding strength compared to H80T20. In literature,[37] it is indicated that an increase in the TiO2 particles
content in HAP matrix leads to the formation of dense coatings with low porosity, thus resulting in coatings with higher hardness and adhesive strength. The H50T50 coating presents approximately a critical load (Lc3) of 5050 mN. With the addition of the SiO2inner
layer on the substrate, the H50T50-S bilayer coating presents a higher bonding strength (Lc4 . 5680 mN).
The SiO2-hydroxylated surface can therefore promote
the attachment of the TiO2/HAP layer on the substrate.
As illustrated in Figure7, such an effect could be due to the formation of Si-O-Ti bridge linking at the interface. The mechanism involved in this interaction is the possibility of the formation of hydrogen bond between the silanol groups, Si-OH, present on the surface of the SiO2film and the Ti-OH or Ti-OR groups Rð ¼ OC3H7Þ bonded to the ends of the alkoxide molecules Ti(OH)yðOC3H7Þ4yafter the hydrolysis of titanium isopropoxide in 2-methoxy ethanol and water,[40] according to the reaction:
Ti OH þ HO Si ! Ti O Si þ H2O
Ti OR þ HO Si ! Ti O Si þ ROH (
½3
This mechanism is similar to condensation reactions taking place during the sol–gel polymerization process as illustrated in Figure7. A similar mechanism has been proposed in the research work of Houmard.[41]
It is reported that the coating thickness should be controlled in such a way as to produce a compromise between the bonding strength and the corrosion prop-erties.[42,43] Therefore, the H50T50-S thickness (1310 nm) obtained in this study was observed to be the optimized coating for improving both the corrosion resistance and the bonding strength.
E. Mechanical Properties
For hardness measurements of hard coatings on soft substrate, some authors[44,45]indicate a value close to 10 pct of the coating thickness for the indenter penetration after which the substrate interferes with the measure-ment. This value can reach 50 pct in the case of soft coating on hard substrate.[46] Unfortunately, this limit which can vary in a great extent is not a predictable value because it depends both on coating thickness and on the mechanical behavior of the coating, i.e., for a hard coating on a soft substrate or for a soft coating on a hard substrate. To avoid the application of models for which the above-mentioned limit values cannot be defined precisely, a direct determination of the mechan-ical properties of the material will be preferable. This is rendered possible by means of the continuous stiffness measurement mode, which allows the computation of both the hardness and the elastic modulus as a function of the indenter penetration.[47] Figure8 represents the hardness variation as a function of the indenter pene-tration. As shown in this figure, H50T50 coating presents the highest hardness value close to the external surface. For the indenter penetrations over than 100 nm, the hardness of H50T50 increases notably to reach 2 GPa, while the hardness of H80T20 coating varies slightly. After 200 nm, the hardness of H50T50 coating decreases and tends to that of H80T20. We note that the hardness value of H80T20 and H50T50 coatings can be considered as a constant value for indenter depths between 500 and 800 nm. In this region, the hardness is close to 1.1 and 1 GPa for the H50T50 and H80T20 coatings, respectively. For penetration depths over 800 nm, the hardness continues to increase to achieve a value which should correspond to that of the substrate (.2 GPa). For the H50T50-S coating, the hardness is rather close to that of H50T50 for indenter penetration close to the outer surface (100 nm), and then, it is clear to note that the hardness is influenced by the presence of the SiO2inner layer. Above 550 nm, the hardness value
tends to that of H80T20 and H50T50 coatings before reaching the substrate hardness over 800 nm in depth. The SiO2 inner layer does not improve the hardness
measurement.
To determine the elastic modulus of H80T20, H50T50, and H50T50-S coatings, a model must be applied for separating the influence of the substrate on the elastic modulus measurement. Indeed, the limit value for the indenter displacement after which the substrate interferes with the measurement is close to 1 pct of the film thickness for a hard film deposited onto a soft substrate[48,49] and it can reach 20 pct for a soft film deposited onto a hard substrate.[46,50] Several models have been proposed to extract intrinsic material prop-erties of the film from the composite modulus, which represent the combination of the respective moduli of the film, Efilm, and of the substrate, Esubstrate. Most of
them are empirical models based on the following relationship:
Ecomposite ¼ Esubstrateþ Eð film EsubstrateÞU xð Þ: ½4 The relative weight U xð Þ of each material, as was pointed out by Doerner and Nix,[50] varies with the penetration depth. Different weight functions are avail-able in literature, and the most commonly used have been proposed by Doerner and Nix.,[51]Mencik et al.,[52] Antunes et al.,[53]and Gao et al.[54]First, we tested the model of Perriot and Barthel[55]who extended the Gao’s function to a larger range of moduli ratios and propose the empirical model in Eq. [5]. This function represents the measured reduced modulus, ERC, as a function of
ERF(obtained for the lowest loads and representing the
reduced modulus of the film) and ERS(obtained for the
highest loads and representing the reduced modulus of the substrate): ERC a t ¼ ERFþ ERS ERF ð Þ 1þ tx0 a n ; ½5
where x0and n are adjustable constants. a (= hÆtanW) is
the contact radius of a assimilated conical indenter at the maximum load, W equals to 70.3 deg corresponding to the half-angle of the tip conical indenter, and h is the indenter displacement and t the film thickness. The parameter x0 is the value of the a/t ratio for which
EC¼ Eð RSþ ERFÞ=2. At the same time, it corresponds to the change in curvature of the (EC; a/t) curve plotted
in semi-log coordinates.
Additionally, we tested the model suggested by Roudet et al.[56] using the mathematical expression similar to that proposed by Avrami[57,58]:
EC¼ EFþ Eð S EFÞexp k½ s:Uns; ½6 where EC is the composite elastic modulus and EF, ES
are, respectively, the elastic modulus of the film and of the substrate. KS, nS are fitting coefficients, which
indi-cate the magnitude of the curve, and the weight func-tion (F) is given by equafunc-tion Eq. [7]:
U¼2 parctan t a þ 1 2p 1ð mcÞ ð1 2mcÞ t aln 1þ a t 2 a t 1þ a t 2 " # : ½7
Finally, Figure9presents the two models applied on the complete range of indentation data. The elastic modulus is represented as a function of the weight function (F) and of the ratio (a/t). The H80T20 (Figure 9(a) and (b)) and H50T50 (Figure9(c) and (d)) curves seem to be satisfactorily represented in all the range of the indentation data. In the range of a/t =00.2 and F =0.91 for the indentation penetration lower than 100 nm, there is some variation in the indentation data related to the insufficient precision of the influence of the indenter tip defect involved in the computation of the contact area. Even the representa-tions are not linear, and the tendencies at the two limits (a/t =0.2) and (F = 0.9) converge toward the elastic moduli of the film. The results of both the elastic modulus of the film and of the substrate as well as the fitting parameters are listed in TableV. The values given by the two models are very similar. For the H50T50-S coating (Figure 9(e) and (f)) when F is lower than 0.7 and the ratio a/t is higher than 1 (for indenter displacements higher than 400 nm), the elastic modulus variation is adequately represented and tends toward that of H50T50 coating (33 GPa). When F is higher than 0.7 and the ratio a/t is lower than 1 (for indenter displacements less than 470 nm), the elastic modulus variation is not adequately represented by the fitting curves. This tendency can be due to the presence of the SiO2inner layer.
The hardness and the elastic modulus of the prepared coatings are found to increase as the content of the TiO2
increases to 50 vol pct. H50T50 possesses the hardness value of 2 GPa at the surface, 1.1 GPa in the core of the coating, and the elastic modulus value of 33 GPa. The hardness variation between the surface and the core of the coating was explained by the initial deposition of some TiO2 on the underlying substrate, while the
remaining TiO2 formed a composite with HAP and
produced the structure.[16]Further deeper analysis about the mechanical properties of the SiO2 film will be
investigated in the future studies in order to explain the influence of the mechanical behavior of the intermediate SiO2layer.
It is noted that the hardness values measured in this study, which are comparable with those reported pre-viously for pure HAP ceramics (1.0-5.5 GPa),[59]are able to withstand large abrasion forces during implant insertion. The elastic modulus has attracted much research interest because of its critical importance for characterizing various bone pathologies and guiding artificial implant design. The equivalent elastic modulus must be adjusted to not greatly exceed those of bones in order to avoid stress shielding at the bone–implant interface, a major source for bone resorption and eventual failure of the implants. The elastic moduli values measured here are comparable with those of cortical bone (compact bone) ranging from 3 to 30 GPa.[60,61]H80T20 and H50T50 coatings are promising materials for the hard tissue application. Owing to its higher corrosion resistance and bonding strength, H50T50-S-coated-316L SS has a potential application in orthopedic as a prosthetic implant. Further studies are required to test the bioactivity and biocompatibility
of these promising biomedical materials in vitro and in vivo. Recently, published work[62]has shown that the TiO2/HAP (20 vol pct TiO2) composite coatings
exhib-ited well biocompatible properties, stem cells attached well onto the surface, proliferated, and presented a polygonal morphology different from the fibroblas-tic-like morphology found on 316L SS. Moreover, TiO2/HAP composite improved the corrosion resistance
of the 316L SS implant and showed mechanical prop-erties close to that of hard tissue once incubated in physiological conditions for 7 days, highlighting its potential application in hard tissue replacement.
IV. CONCLUSION
The 50 vol pct TiO2-reinforced HAP layer has been
successfully attached on the surface of silica-coated 316L SS by sol–gel method. The surface properties of silica inner layer enhanced the attachment of the TiO2
and HAP phases through the Si-OH hydroxyl groups and the chemical affinity of TiO2toward the HAP, thus
suggesting an improvement in the corrosion resistance, bonding strength, elastic modulus, and hardness of the coated 316L SS required for the hard tissue application.
ACKNOWLEDGMENTS
This study was funded by the University of Bejaia from Algeria. The authors are very grateful to MSMP, Arts et Me´tiers ParisTech of Lille, France, for the help in providing the mechanical analysis.
REFERENCES
1. E.C. Shores, R.E. Holmes: An Introduction to Bioceramics, in: L.L. Hench, J. Wilson (Eds.), World Scientific: Singapore, 1993, pp. 181–98.
2. Marc Long, H.J. Rack: Biomaterials, 1998, vol. 19, pp. 1621–39. 3. M. Navarro, A. Michiardi, and O. Castan˜o: J.A PlanellJ. R. Soc.
Interface, 2008, vol. 5, pp. 1137–58.
4. H. Hermawan, D. Ramdan, J.R.P. Djuansjah: Metals for Biomedical Applications, Biomedical Engineering - From Theory to Applications, Prof. Reza Fazel (Ed.), InTech, Available from:
http://www.intechopen.com/books/biomedical-engineering-from-theory-toapplications/, 2011.
5. V.P. Mantripragada, B. Lecka-Czernik, N.A. Ebraheim, and A.C. Jayasuriya: J Biomed Mater Res A, 2013, vol. 101, pp. 3349–64. 6. M. Navarro, A. Michiardi, O. Castano, and J. Planell: J. R. Soc.
Interface, 2008, vol. 5, pp. 1137–58.
7. J.A. Disegi and L. Eschbach: Injury, 2000, vol. 31, pp. 2–6. 8. J. Blumn, K.L. Eckert, A. Schroeder, M. Petitmermet, S.W. Ha, E.
Wintermantel: Proceedings of the 9th International Symposium on Ceramics in Medicine, in: T. Kokubo, T. Nakamura, F. Miyaji (Eds.), Col. 9, Otsu, Japan, 1996, pp. 89.
9. S. Nagarajan and N. Rajendran: Appl. Surf. Sci., 2009, vol. 255, pp. 3927–32.
10. H.W. Kim, Y.H. Koh, L.H. Li, S. Lee, and H.E. Kim: Biomate-rials, 2004, vol. 25, pp. 2533–38.
11. C.E. Wen, W. Xu, W.Y. Hu, and P.D. Hodgson: Acta Biomater., 2007, vol. 3, pp. 403–10.
12. T.P. Singh: Harpreet SinghHazoor Singh: J. Therm. Spray Tech-nol., 2012, vol. 21, pp. 917–27.
13. P.C. Rath, L. Besra, B.P. Singh, and S. Bhattacharjee: Ceram. Inter., 2012, vol. 38, pp. 3209–16.
14. D. Sidane, D. Chicot, S. Yala, F. Bir, H. Khireddine, S. Ziani, A. Iost, and X. Decoopman: Thin Solid Films, 2015, vol. 593, pp. 71–80.
15. A.J. Nathanael, N. Sabari Arul, N. Ponpandian, D. Mangalaraj, and P.C. Chen: Thin Solid Films, 2010, vol. 518, pp. 7333–38. 16. A.J. Nathanael, D. Mangalaraj, and N. Ponpandian: Compos. Sci.
Technol., 2010, vol. 70, pp. 1645–51.
17. H. Li, K.A. Khor, and P. Cheang: Biomaterials, 2003, vol. 24, pp. 949–57.
18. P. Li, C. Ohtsuki, T. Kokubo, K. Nakanishi, N. Soga, T. Nakamura, and T. Yamamuro: J. Am. Ceram. Soc, 1992, vol. 75, pp. 2094–97.
19. P. Li, C. Ohtsuki, T. Kokubo, K. Nakanishi, N. Soga, and K. de Groot: J Biomed Mater Res, 1994, vol. 28, pp. 7–15.
20. P. Galliano, J.J. Damborenea, M.J. Pascual, and A. Duran: J. Sol-Gel Sci. Technol., 1998, vol. 13, pp. 723–27.
21. S. Kim and P.N. Kumta: Mater. Sci. Eng. B, 2004, vol. 111, pp. 232–36.
22. A. Balamurugan, S. Kannan, and S. Rajeswari: Mater. Lett., 2005, vol. 59, pp. 3138–43.
23. T. Kokubo and H. Takadama: Biomaterials, 2006, vol. 27, pp. 2907–15.
24. J.E. Field and R.H. Telling: The young modulus and poisson ratio of diamond, Research Note, Cavendish Laboratory, Cambridge, 1999. 25. M. Houmard, E.H.M. Nunes, D.C.L. Vasconcelos, G. Berthome´, J.-C. Joud, M. Langlet, and W.L. Vasconcelos: Appl. Surf. Sci., 2014, vol. 289, pp. 218–23.
26. S. Permpoon, G. Berthome´, B. Baroux, J.C. Joud, and M. Langlet: J. Mater Sci., 2006, vol. 41, pp. 7650–62.
27. L. Zhang, Z. Xing, H. Zhang, Z. Li, X. Wu, X. Zhang, Y. Zhang, and W. Zhou: Appl Catal B., 2016, vol. 180, pp. 521–29. 28. P.K. Jal, M. Sudarshan, A. Saha, S. Patel and B.K. Mishra:
Col-loids Surf. A: Chem. Phys. Eng. Asp., 2004, vol. 240, pp. 173–78. 29. V.A. Ermoshin, K.S. Smirnov, and D. Bougeard: Surf. Sci., 1996,
vol. 368, pp. 147–51.
30. N. Primeau, C. Vautey, and M. Langlet: Thin Solid Films, 1997, vol. 310, pp. 47–56.
31. M. Houmard, D. Riassett, F. Roussel, A. Bourgeois, G. Berthome´, J.C. Joud, and M. Langlet: Surf. Sci., 2008, vol. 602, pp. 3364–74.
32. D. Sidane, H. Khireddine, S. Yala, S. Ziani, F. Bir, D. Chicot: Metall. Mater. Trans. B, 2015, vol. 46 B, pp. 2340–47.
33. C. Ergun: J Eur Ceram Soc, 2008, vol. 28, pp. 2137–49. 34. C.C. Ribeiro, I. Gibson, and M.A. Barbosa: Biomaterials, 2006,
vol. 27, pp. 1749–61.
35. H. Anmin, L. Ming, C. Chengkang, and M. Dali: J. Mol. Catal. A: Chem., 2007, vol. 26, pp. 779–85.
36. S. Yala, H. Khireddine, D. Sidane, S. Ziani, and F. Bir: J. Mater. Sci., 2013, vol. 48, pp. 7215–23.
37. S. Yugeswaran, A. Kobayashi, A. Hikmet Ucisik, and B. Subramanian: Appl. Surf. Sci., 2015, vol. 347, pp. 48–56.
38. Y. Huang, Q. Ding, X. Panga, and S. Hana: Y. YanAppl. Surf. Sci., 2013, vol. 282, pp. 456–62.
39. A.S. Toloei, A. Stoilov, and D.O. Northwood: WIT Transactions on Engineering Science, 2013, vol. 77, pp. 193–204.
40. A. Balamurugan, S. Kannan, and S. Rajeswari: Mater. Lett., 2005, vol. 59, pp. 3138–43.
41. M. Houmard: PhD. Thesis, INP Grenoble, France, 2009. 42. B. Aksakal, M. Gavgali, and B. Dikici: J. Mater. Eng. Perform,
2010, vol. 19, pp. 894–99.
43. A. Bu¨yu¨ksagisa, N. C¸iftc¸i, Y. Ergu¨n, and Y. Kayali: Prot. Met. Phys. Chem., 2011, vol. 47, pp. 670–79.
44. H. Buckle:The Science of hardness testing and its research appli-cationsin ASM, J.W. Westbrook and H. Conrad, eds., Metals Park, OH, 1973, pp. 453–91.
45. Y. Sun, T. Bell, and S. Zheng: Thin Solid Films, 1995, vol. 258, pp. 198–204.
46. Z.H. Xu and D. Rowcliffe: Thin Solid Films, 2004, vols. 447–448, pp. 399–405.
47. W.C. Oliver and G.M. Pharr: J. Mater. Res, 1992, vol. 7, pp. 1564–83.
48. T. Chudoba, N. Schwarzer, and F. Richter: Surf. Coat. Technol., 2002, vol. 154, pp. 140–51.
49. F. Cleymand, O. Ferry, R. Kouitat, A. Billard, and J. von Stebut: Surf. Coat. Technol., 2005, vol. 200, pp. 890–93.
50. T. Ohmura, S. Matsuoka, K. Tanaka, and T. Yoshida: Thin Solid Films, 2001, vol. 385, pp. 198–204.
51. M.F. Doerner and W.D. Nix: J. Mater. Res., 1986, vol. 1, pp. 601–09.
52. J. Mencik, D. Munz, E. Quandt, E.R. Weppelmann, and M.V. Swain: J. Mater. Res., 1997, vol. 12, pp. 2475–92.
53. J.M. Antunes, J.V. Fernandes, N.A. Sakharova, M.C. Oliveira, and L.F. Menezes: Int. J. Solids Struct., 2007, vol. 44, pp. 8313–34. 54. H. Gao, C.H. Chiu, and J. Lee: Int. J. Solids. Struct., 1992, vol. 29,
pp. 2471–92.
55. A. Perriot and E. Barthel: J. Mater. Res., 2004, vol. 19 (02), pp. 600–08.
56. F. Roudet, D. Chicot, X. Decoopmana, A. Iost, J. Bu¨rgi, J. Garcia-Molleja, L. Nosei, and J. Feugeas: Thin Solid Films, 2015, vol. 594, pp. 129–37.
57. M. Avrami: J. Chem. Phys., 1940, vol. 8, pp. 212–24. 58. M. Avrami: J. Chem. Phys., 1941, vol. 9, pp. 177–84.
59. J. Song, Y. Liu, Y. Zhang, and L. Jiao: Mater. Sci. Eng. A, 2011, vol. 528, pp. 5421–27.
60. X. Wang, S. Xu, S. Zhou, W. Xu, M. Leary, P. Choong, M. Qian, M. Brandt, and Y. Min: Xie: A reviewBiomaterials, 2016, vol. 83, pp. 127–41.
61. A.J. Nathanael, Y.M. Ima, T.H. Oha, R. Yuvakkumarb, and D. Mangalarajc: Appl. Surf. Sci., 2015, vol. 332, pp. 368–78. 62. D. Sidane, H. Rammal, A. Beljebbar, S.C. Gangloff, D. Chicot, F.
Velard, H. Khireddine, A. Montagne, and H. Kerdjoudj: Mater. Sci. Eng. C, 2017, vol. 72, pp. 650–58.
